Individuals with autism spectrum disorder (ASD) exhibit differences in attentional patterns from very early in life, and these attentional differences have been implicated in the developmental cascade that gives rise to core features of the disorder such as language and social communication deficits. Previous research in ASD suggests that attention differences are most prominent for dynamic social stimuli, and it has been hypothesized that the complexity and relative unpredictability of social interactions underlie these social attention deficits. Substantial evidence shows that typically developing (TD) children attend to and extract probabilistic relationships between events in the environment, and that from infancy, they prefer attending to intermediately predictable contingencies. Attending to these intermediately complex probabilistic relationships allows for learning of developmentally appropriate information?not so simple that it is already known, and not so complex that it is impossible to learn. Clinical and experimental evidence suggests that children with ASD prefer highly predictable events and stimuli. A preference for overly predictable stimuli may lead children with ASD to select developmentally atypical stimuli for processing and miss out on critical opportunities for learning. The current proposal seeks to quantify the impact of stimulus predictability on attentional preferences in children with ASD as compared to TD peers across two key learning contexts?observing and interacting with the environment. Specifically, this project will use eye tracking to examine how attention is impacted by stimulus predictability in a passive observation context (Aim 1). Additionally, this project will also examine how attentional preferences are impacted by predictability in children?s interactions with the environment via a freeplay touch screen task (Aim 2). These aims will be tested using computational modeling to quantify and compare attentional preferences in both learning contexts between young children (ages 4-6) with ASD (n=35) and nonverbal mental age matched TD controls (n=35). The proposed project will advance the field by examining a general cognitive processing style that has particular implications for the ability to attend to and learn from complex social stimuli, a key developmental process. Understanding how children with ASD select information for processing can provide a unifying understanding of the development of core deficits in ASD across both social communication and restricted/repetitive behavior symptom domains. Finally, understanding early attention differences has the potential to inform markers for early identification and targets for intervention in ASD.